Study Design. (A) The initial study population consisted of CAMP probands used in (1) a case-control design composed of 359 CAMP cases and 846 Illumina controls and (2) a family-based design composed of 403 trios. Genome-wide association of individual SNPs to asthma status was assessed in each of these designs. SNPs with a case-control Cochran-Armitage trend test p-value < 0.01 and a family-based PBAT additive model p-value < 0.05 were selected for replication analysis. (B) The first replication stage, carried out in CR, measured the association of 1378 SNPs with asthma. Those SNPs with either (1) a CR PBAT additive model p-value < 0.05 or (2) a CR PBAT additive model p-value < 0.10 and a p-value < 1E-05 from the designs in (A) were selected for the next replication stage. (C) The second replication stage was carried out in two additional independent populations, FHS and iCAP. (D) Joint association analysis for 85 SNPs with data in the three independent populations was performed.